Automated trust-and-safety at scale: classifiers, deepfake detection and policy enforcement.
266 articles
LaLiga's aggressive efforts to block illegal sports streaming are inadvertently taking down legitimate websites. This highlights the risks of automated IP blocking systems in large-scale content protection.
Bloomberg Technology · 81d ago
A recent report indicates that NSFW-related queries account for over half of all traffic on the Grok platform. This highlights a significant trend in user behavior regarding unfiltered AI models.
Engadget · 86d ago
A new report by Chayn argues that tech companies and authorities are misdirecting their efforts by focusing solely on nudity. The report calls for a shift toward prioritizing consent in the fight against online abuse.
BBC Technology · 90d ago
A new study reveals that AI-generated junk content has become a defining feature of TikTok. The platform's younger users and educational channels are disproportionately affected by this influx of low-quality automated media.
Digital Trends · 90d ago
UK regulator Ofcom has contacted Telegram to demand clarification on how the platform detects and prevents illegal incitement. This follows the use of the app by a handler to coordinate arson attacks on properties linked to Keir Starmer.
The Guardian Technology · 92d ago
Telegram has filed a petition in the Delhi High Court to overturn an Indian government order blocking the app. The block is reportedly linked to investigations into cheating on a national medical-entrance exam.
The Next Web (TNW) · 94d ago
This article explores the implications of specific prompts that lead ChatGPT to generate disturbing or inappropriate imagery. It raises critical questions about AI safety, guardrails, and the challenges of content moderation in generative models.
BBC Technology · 95d ago
Following a US Senate investigation, Spotify removed 57,000 AI-generated podcast episodes promoting illegal drugs. The platform also banned 3,500 accounts linked to these illicit activities.
The Next Web (TNW) · 97d ago
X has faced backlash for refusing to remove hate speech, including racial slurs directed at prominent UK politicians. Despite reports from researchers, the platform failed to take action, raising concerns about its automated and manual moderation efficacy.
The Guardian Technology · 97d ago
Spotify has removed over 57,000 podcast episodes following a congressional report identifying them as vehicles for promoting illegal pharmacy websites. These fake podcasts were used to manipulate search rankings and drive traffic to illicit online drug marketplaces.
Digital Trends · 98d ago
Reports indicate that a neo-Nazi youth group utilized social media anger, fueled by figures like Elon Musk, to orchestrate anti-immigrant riots in Belfast. This highlights the role of algorithmic amplification in real-world civil unrest.
Wired · 99d ago
Senators Cruz and Wyden have introduced the JAWBONE Act, a bipartisan legislative effort aimed at addressing government-led censorship. The bill seeks to establish clearer boundaries regarding government influence over digital content moderation.
Engadget · 99d ago
Polymarket and Kalshi have instructed influencer partners to remove social media content that questions the integrity of election results. These platforms are tightening oversight on paid partnerships to prevent the spread of misinformation regarding democratic processes.
Wired · 103d ago
Ofcom has issued a formal warning to social media platforms to implement robust measures against illegal hate content during the World Cup. The regulator will actively monitor platform performance to ensure the safety of minority ethnic players.
The Guardian Technology · 103d ago
Meta's AI-driven feed is reportedly struggling with an influx of low-quality, engagement-driven content. The platform is seeing a rise in fake stories and clickbait as it pushes for a social AI experience.
Digital Trends · 105d ago
Guy Rosen, Meta’s Chief Information Security Officer who previously led election integrity initiatives, is set to leave the company in the coming months. His departure marks a significant leadership change within Meta's safety and security division.
Bloomberg Technology · 109d ago
Meta is implementing new safety features to restrict harmful content exposure for teenagers on Instagram, Facebook, and Messenger. These updates follow recent legal challenges regarding child safety and platform responsibility.
New York Times Technology · 109d ago
Instagram is implementing new limitations on the types of content teens can repeatedly view. These restrictions specifically target posts related to body image and mental health to improve user safety.
Engadget · 109d ago
The UK government is considering legislation to restrict social interactions for minors under 16 in popular games like Fortnite and Minecraft. This move aims to enhance child safety by regulating communication features on gaming platforms.
cnBeta (Full RSS) · 110d ago
Fanqie Novel has intensified its governance of low-quality content, disposing of over 150,000 books in May. This includes rejecting new contracts and removing existing non-compliant content.
36氪 · 117d ago
The automotive industry faces a surge in organized, AI-generated smear campaigns that exploit high-stakes consumer decision-making. Despite legal victories by car manufacturers, the low cost of AI-powered content production and platform traffic incentives fuel this gray-market industry.
虎嗅 · 122d ago
A prestigious short story prize winner is under scrutiny after AI detection tools and syntactical analysis raised concerns about potential AI authorship. Both the awarding foundation and the publisher have yet to reach a definitive conclusion regarding the work's origin.
The Guardian Technology · 123d ago
Instagram plans to penalize 'unoriginal' posts to improve feed quality. This targets 'feed slop,' low-quality content overwhelming user timelines.
Engadget · 141d ago
Spotify is introducing 'Verified' badges to profiles to differentiate human artists from AI-generated ones. The platform will assess criteria like live performance dates and social media presence.
BBC Technology · 141d ago
Spotify launched a 'Verified by Spotify' green checkmark for artist profiles to confirm human authenticity amid surging AI-generated tracks. The badge will roll out on profiles and search results after platform review.
The Guardian Technology · 142d ago
Meta terminated Kenyan contractors who reported witnessing users engaging in sexual activities through Ray-Ban Meta smart glasses. The company stated these workers did not meet their standards.
Ars Technica AI · 142d ago
Divine, a reboot of the Vine short-video app, has launched featuring a strict ban on AI-generated 'slop' content. The app aims to revive authentic, human-created videos.
Engadget · 143d ago
Two California Democrats introduced a bill to allow lawsuits against social media companies for failing to detect or remove child sexual abuse material. Lawmakers Maggy Krell and Buffy Wicks cite profound trauma from online exploitation.
The Guardian Technology · 148d ago
Child safety officers need larger budgets to protect kids from online abuse. Easier-to-use AI tools have caused a rise in abusive imagery.
Bloomberg Technology · 149d ago
Generative AI is rapidly increasing child sexual abuse imagery production. Detection and policing efforts by watchdogs are overwhelmed.
Digital Trends · 149d ago